Attempts to emulate the serendipity and subsequent rapport that occurs between mentors and mentees in informal mentoring relationships frequently challenges the organizers of many formal mentoring schemes. Using qualitative evidence from a community mentoring project, research is presented that suggests that, through careful mentor selection and appropriate training, the matching of mentors and mentees, except perhaps by geographical location and time availability, is unnecessary. Often, as is shown here, it is not until some way into the relationship that totally unanticipated coincidences between mentor and mentee are discovered; and often these associations are ones that mentoring scheme organizers could not have anticipated. (Contains 3 figures.)